4 ms·

>I also push often because I'm forever aware disks can fail. In the 20+ years that I've been using computers, and ≈15 or so that I've been writing software, I'

>I also push often because I'm forever aware disks can fail. In the 20+ years that I've been using computers, and ≈15 or so that I've been writing software, I've never experienced a drive failure.

secondcoming 5y ago
I have. You're not coding hard enough!
El_RIDO 5y ago
This is a good example of how scale can bias your view! :-D Personally I share the same experience: In the 30 years I've been using personal computers at home and work, only 2 disks failed on me, one was an error on my part during an OS upgrade and the second was an external disk that physically fell off my bed while I had it plugged into my notebook. So both of these weren't really those random failures. On the other hand, while working in an IT infra team of a 60 developer company with 4 racks full of servers for 3 years, we'd get to see about one failed SSD per 2 months and 2-3 failed disks or SSDs per year in servers. During the 1 year I worked for the IT infra team at a smaller hospital of a medium sized city with a large VDI environment and two HP EVAs in a two datacenter configuration each, we'd get 3-4 disk failures per week. Those had over 140 disk per storage each and were approaching the end of their support life, being around 5 years old, so the failure ratio started to get higher.
amarshall 5y ago
Sure, but your anecdote does not refute in any way that they can fail. It’s all about risk vs. overhead of avoiding that risk. The overhead of pushing often is very low, so even with the low probability of a drive failure, there’s not much reason to not do so (or any other form of backup, really).
TillE 5y ago
I've had a Macbook SSD fail (suddenly and completely) within the warranty period. Surprisingly I've only had one hard drive fail before I retired it, maybe when it was about 6 years old. But those are just the catastrophic failures, I'm pretty sure I've also had hardware-related data corruption here and there as well.
contingencies 5y ago
Some non physical failure scenarios I have experienced include catastrophic filesystem failures (NTFS: never again), loss of machine and no handy backup machine capable of reading the disk or filesystem in question (extX, ZFS, etc.), accidental damage to disk, partition or filesystem tree due to bad code (operator error; PEBSAC).
brimble 5y ago
About 30 years using computers, a little over 20 being paid to do it. I've had two fail outright (one internal, one external) and a couple others develop errors that I noticed (one because I finally started using ZFS on my bulk-storage machine and it's very good at telling when a drive is misbehaving, the other, SMART caught, years ago) so they had to be replaced. All were spinning rust. Maybe 50 drives total, over my computer-using life, counting those in game consoles and ones in work machines that were assigned to me. Including SSDs. Give or take 10. With the experience of the last 3 or so years of using ZFS, I'd bet most of my "still working" drives over the years were erroring and losing data by the time I replaced them, and I just didn't know it.

>I also push often because I'm forever aware disks can fail. In the 20+ years that I've been using computers, and ≈15 or so that I've been writing software, I've never experienced a drive failure.

secondcoming 5y ago

I have. You're not coding hard enough!

El_RIDO 5y ago

This is a good example of how scale can bias your view! :-D Personally I share the same experience: In the 30 years I've been using personal computers at home and work, only 2 disks failed on me, one was an error on my part during an OS upgrade and the second was an external disk that physically fell off my bed while I had it plugged into my notebook. So both of these weren't really those random failures. On the other hand, while working in an IT infra team of a 60 developer company with 4 racks full of servers for 3 years, we'd get to see about one failed SSD per 2 months and 2-3 failed disks or SSDs per year in servers. During the 1 year I worked for the IT infra team at a smaller hospital of a medium sized city with a large VDI environment and two HP EVAs in a two datacenter configuration each, we'd get 3-4 disk failures per week. Those had over 140 disk per storage each and were approaching the end of their support life, being around 5 years old, so the failure ratio started to get higher.

amarshall 5y ago

Sure, but your anecdote does not refute in any way that they can fail. It’s all about risk vs. overhead of avoiding that risk. The overhead of pushing often is very low, so even with the low probability of a drive failure, there’s not much reason to not do so (or any other form of backup, really).

TillE 5y ago

I've had a Macbook SSD fail (suddenly and completely) within the warranty period. Surprisingly I've only had one hard drive fail before I retired it, maybe when it was about 6 years old. But those are just the catastrophic failures, I'm pretty sure I've also had hardware-related data corruption here and there as well.

contingencies 5y ago

Some non physical failure scenarios I have experienced include catastrophic filesystem failures (NTFS: never again), loss of machine and no handy backup machine capable of reading the disk or filesystem in question (extX, ZFS, etc.), accidental damage to disk, partition or filesystem tree due to bad code (operator error; PEBSAC).

brimble 5y ago

About 30 years using computers, a little over 20 being paid to do it. I've had two fail outright (one internal, one external) and a couple others develop errors that I noticed (one because I finally started using ZFS on my bulk-storage machine and it's very good at telling when a drive is misbehaving, the other, SMART caught, years ago) so they had to be replaced. All were spinning rust. Maybe 50 drives total, over my computer-using life, counting those in game consoles and ones in work machines that were assigned to me. Including SSDs. Give or take 10. With the experience of the last 3 or so years of using ZFS, I'd bet most of my "still working" drives over the years were erroring and losing data by the time I replaced them, and I just didn't know it.